When an explainer video feels cheap and you can't say why, the problem is usually timing: two related things changing at unrelated moments, a frame frozen while the voice keeps talking, a cursor gliding at a speed no hand has ever moved. Across our produced videos, most review notes have been timing notes. Reviewers rarely say a box is the wrong shade of blue. They say "the clicks appear too quickly," or "this part just sits there," or just "boring."
Picture an uptime monitor: it pings your services on a schedule, opens an incident when a check fails, posts the incident to a timeline, and pages whoever is on call. A video about that product is a chain of causes from its first frame.
Viewers read timing as causality
A video has no narrator's finger on the screen and no arrows anyone actually reads. What the viewer has is order in time: when two things change together, the brain files them as connected, and when they change at unrelated moments, it files them as strangers. That happens whether you planned it or not, so every timing decision changes what viewers believe caused what.
In one early build, a query block ran, and then, on its own schedule, some table rows highlighted. Both events happened, and nobody watching connected them. We re-timed the highlights to land against the block going live, and the same pixels started explaining themselves.
In the uptime video, a check fails and the incident row fills a beat later, so the viewer reads the failure as cause and the incident as consequence. The beat matters as much as the order — cause should lead effect by about 0.7 seconds (where that number comes from), because simultaneous events read as coincidence and a long gap snaps the connection — which sets two rules:
- Give every animation a named sync partner. An animation that cannot name the other thing it changes with is decoration, and decoration is a large part of why videos look cheap.
- Let cause lead effect by a short, fixed beat. The check turns red, the incident row fills a moment later, every time at the same offset. The consistency lets the viewer trust the grammar.
Dead holds are the most common failure
A hold is a stretch where the layout rests. A hold after a payoff gives the viewer a moment to absorb it. The failure is the dead hold: a frame settles early and sits frozen while the voiceover finishes, and viewers feel the video stop caring even when they can't say why.
We measured one of our own bad builds: 44% of its runtime was fully static, which is 52 seconds of settled frames in a 93-second video. Another build ended every one of its six scenes with three to five seconds of stillness. It passed every mechanical check we had, and it died in review anyway.
Pause the video at random, several times. If you land on a completely still frame more than a third of the time, you have made a slideshow with a soundtrack. Three rules fix it:
- Cap dead holds at about 3 seconds. Even that long a hold is only earned after something has happened. Before anything happens, don't hold at all.
- Budget ambient life into every hold longer than a second. A slow drift, a pulsing status light, a working dot on the check that is currently probing.
- End scenes on the question, not the answer. Hold on the check still burning red before the page goes out; never camp on the all-green board.
The easing window matters more than the curve
Easing is the shape of a movement over time. Every animation tool offers a menu of curves, and the menu is where most people stop thinking. One of our builds showed why that fails: a cursor moved between two click targets on a textbook ease-in-out, but the clicks were five seconds apart in the script and the movement had been stretched across the whole gap, so the cursor glided for 4.5 seconds. Review flagged it instantly. Nothing about the curve was wrong; it just didn't move like a hand.
The fix was a different window, not a different curve. A real hand dwells at the last target, makes one quick flight with a fast launch and a long deceleration, and settles a beat before the click. Rebuilt that way — flights of 0.32 to 0.55 seconds scaled to distance, settling about 0.12 seconds before the click — the cursor read as a person. Anything imitating physical behavior takes its duration from the behavior and only its shape from the curve menu. Four defaults follow:
- Arrivals decelerate. Nothing physical arrives at full speed.
- Camera moves ease on both ends. The camera is a head turning, and heads don't snap.
- Physical-feeling elements get a small spring. A tiny overshoot reads as weight.
- Decide the profiles once. Re-tuning easing by feel on every project produces a library of videos with a different accent in each one.
Rhythm is the spacing between events
Uniform timing reads as a slideshow. Real systems don't finish their work in order at even intervals, and viewers know it in their bones: twelve checks completing at exactly one-second gaps would look like a screensaver. Three tools put the rhythm back:
- Stagger with a deliberate cadence. About 0.35 seconds per item reads as a narrator pointing at things one by one; about 0.14 reads as a machine sweeping a selection. When the voice walks through five checks by name, light them at 0.35; when the monitor sweeps a region into maintenance, run it at 0.14. Swap them and the machine looks sluggish or the narrator frantic.
- Scramble the finish order, keep the gaps even. Shuffle the order results land in, and lock the spacing so the sequence stays readable. Randomize both and it's chaos. Regularize both and it's a slideshow again.
- Compress gaps to show load. Incidents arriving 1.6 seconds apart, then 1.4, then 1.2 read as a system under pressure. Density comes from overlap, never from faster elements.
The beat shape also has to keep changing. One of our technically cleanest builds repeated the same well-made beat five times over the same two surfaces, and the entire review verdict was one word: "boring." The builds that graded best changed shape as they went — a wide pull, a fan-out, a close-up, a scrambled finish. Where the camera fits into that variety is its own craft.
When uniform timing is the right choice
When the message is simultaneity, uniform is honest: if the point is that all five regions update together the moment a config changes, together is how they should move. Let sameness mean "these are the same" rather than "we didn't think about it."
Timing belongs in the plan, not the polish pass, so we settle sync partners, holds, easing windows, and cadences at the storyboard stage, before a single frame renders.
FAQ
Do I need to know animation software to use any of this? No. Each rule is a note about when something happens: "incident row fills a beat after the check fails," or "hold 2 seconds, keep the probe dot spinning." Whoever animates it executes better with those notes than without.
What's the fastest way to check a finished video for timing problems? Scrub to random frames — repeated stills mean dead holds. Then watch it muted — if you can't tell what caused what, the sync is broken.
Is slower animation more premium? Appropriate is premium. A 4.5-second cursor glide reads as broken, not luxurious. Slow suits camera moves between ideas; clicks and arrivals should move at the speed the real behavior moves.
How exact do these numbers need to be? Treat them as strong defaults. All of them came from graded work, and what matters most is picking values deliberately and keeping them consistent across the video.
If you want this done for your product, start with the URL. We come back with twenty short candidates and you pick from the board.